Patent: defect early warning system

The invention provides a defect early warning method and system, which relates to the field of Smart Manufacturing. A defect early warning method includes: data filtering and data quality optimization, wherein, filtering out wrong data and modifying missing and / or wrong data; Select the parameter series that affect the defective products in the MES database, and then carry out engineering modeling and machine learning based on the selected data to realize the close correlation between the model and the data and improve the high accuracy of the model; Before the production of the product is completed, judge whether the product will have defects for the first time. If the judgment result is that there will be defects, alarm and change the parameter combination or even replace the worn parts. When the product is not processed, it can predict whether the product is genuine or defective after production based on historical data and processing data of the product. In addition, the invention also provides a defect early warning system, which comprises a data processing module, a modeling module and a defect early warning module.

A Defect Warning Method and System

The invention relates to the field of Smart Manufacturing, in particular to a defect early warning method and system.

Background Technology

At present, most project management systems do not have intuitive and effective prediction methods for possible defects of products. When there are test defects, developers cannot quickly obtain test defects. Defect early warning system is of great significance in manufacturing industry with generally low product quality and serious defects!

At present, all MES have a common weakness, that is, MES can only show the quality problems and defective products of products at all stages, but it does not have the intelligence to automatically provide optimization schemes to turn the products that would otherwise be defective into genuine products, or provide alarms before production.

For example, the defective rate of lithium battery manufacturing industry has been claimed by South Korea and Japan to be less than 2%, while China has been close to 10% for many years. How to make the defect early warning system simple, easy to use and modular is an urgent problem for technicians in this field.

Summary Of The Invention

The purpose of the invention is to provide a defect early warning method, which can predict whether the product is genuine or defective after completion of production based on historical data and processing data of the product when the product is not processed; If it will be a defective product, change the combination of production process parameters or even change the worn parts when the production is not completed, so that the product becomes a genuine product. This system uses the data of MES, the industrial Internet and the test data of the factory test department as the data source, and does not need to collect data to cause product loss, so the development cost is relatively low.

Another object of the invention is to provide a defect early warning system, which can run a defect early warning method.

The embodiment of the invention is realized as follows:

In the first aspect, the embodiment of the application provides a defect early warning method, which includes data filtering and data quality optimization, in which the wrong data is filtered and the missing and / or wrong data is modified; Select the parameter series that affect the defective products in the MES database, and then carry out engineering modeling and machine learning based on the selected data to realize the close correlation between the model and the data and improve the high accuracy of the model; Before the production of the product is completed, judge whether the product will have defects for the first time. If the judgment result is that there will be defects, alarm and change the parameter combination.

In some embodiments of the invention, before the production of the product is completed, judge whether the product will have defects for the first time. If the judgment result is that there will be defects, alarm and change the parameter combination. After that, it also includes: judge whether the product will have defects for the second time after changing the parameter combination, and replace the wear parts if the judgment result is that there will be defects.

In some embodiments of the invention, data filtering and data quality optimization are carried out as described above, wherein filtering out wrong data and modifying missing and / or wrong data include: product quality defect detection based on machine vision image recognition model, image recognition and data analysis based on the perception layer in the image recognition model, So as to obtain the data and filter and optimize the quality of the obtained data.

In some embodiments of the invention, the above selects the parameter series affecting this defective product in the MES database, and then carries out engineering modeling and machine learning based on the selected data to realize the close correlation between the model and the data and improve the high accuracy of the model, including establishing an artificial intelligence algorithm based on industrial scene for real-time data access processing, model calculation Engineering model of rule judgment and real-time early warning.

In some embodiments of the invention, the above selects the parameter series affecting the defective product in the MES database, and then carries out engineering modeling and machine learning based on the selected data to realize the close correlation between the model and the data and improve the high accuracy of the model, including preprocessing the parameters affecting the defective product in the MES database, wherein, Preprocessing includes mean denoising, brightening the area to be identified by Laplace operator, and extracting the possible defect data by gray feature.

In some embodiments of the invention, before judging whether the product will have defects for the first time before the production of the product is completed, if the judgment result is that there will be defects, give an alarm and change the parameter combination, it also includes optimizing the judgment rules through deep learning and establishing a defect judgment rule base based on machine learning.

In some embodiments of the invention, the above also includes: when the preset percentage value of the maximum allowable defect is reached, the alarm is turned on to prevent the product from becoming defective.

In the second aspect, the embodiment of the application provides a defect early warning system, which includes a data processing module for data filtering and data quality optimization, in which the wrong data is filtered and the missing and / or wrong data is modified;

The modeling module is used to select the parameter series affecting this defective product in the MES database, and then carry out engineering modeling and machine learning based on the selected data, so as to realize the close correlation between the model and the data and improve the high accuracy of the model;
The defect early warning module is used to judge whether the product will have defects for the first time before the production of the product is completed. If the judgment result is that there will be defects, it will alarm and change the parameter combination.

In some embodiments of the invention, the above includes: at least one memory for storing computer instructions; At least one processor communicating with the memory, wherein when the at least one processor executes the computer instruction, the at least one processor causes the system to execute: data processing module, modeling module and defect early warning module.

Third, the embodiment of the present application provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, it implements the method of any one of a defect early warning method.

Compared with the prior art, the embodiment of the invention has at least the following advantages or beneficial effects:

First, model the influencing factors of the defects that need early warning, and improve its accuracy through machine learning. If the model predicts that the products currently being produced will be defective after production, the system will give an alarm to prompt the on-site operators to change the combination of process parameters, or even replace the worn parts in advance, so that the products being produced, It will be genuine after production in the future;

Second, when the operator carries out the best combination of process parameters after receiving the alarm, the system provides the reference parameters of the best combination, which can be used by the operator. The second main function is to set the best parameters by yourself, and automatically provide the parameter combination for the operator's reference.

There are many function points in each of the two main functional areas. When setting the early warning parameters, three coefficients can be added to eliminate the model error, and the coefficients during the initial alarm can also be set.